Apparently many DACs that use I2S architecture and related chipsets do not correctly convert 16 bit samples to 24 or 32 bit words prior to conversion. What measures are typically taken to correct this problem, as described below? Which chipsets and DAC architectures are more subject to it...